Nalco_corrosion_manual_72.pdf

Corrosion in the Petroleum Industry

TABLE OF CONTENTS I.

CARBON DIOXIDE CORROSION CHARACTERISTICS "SWEET CORROSION" . . . . . . . . . . . . . . . . . . . . . . . . 1 CARBON DIOXIDE CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 II. OXYGEN CORROSION CHARACTERISTICS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 OXYGEN CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 III. HYDROGEN SULFIDE CORROSION CHARACTERISTICS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 HYDROGEN SULFIDE CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 IV. MICROORGANISM INFLUENCED CORROSION CHARACTERISTICS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 MICROORGANISM INFLUENCED CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 OBTAINING AND ENUMERATING A SAMPLE OF SESSILE BACTERIA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 V. HYDROCHLORIC ACID (STRONG ACID) CORROSION CHARACTERISTICS . . . . . . . . . . . . . . . . . . . . . . . . 55 HYDROCHLORIC ACID CORROSION — LOW PH AQUEOUS CORROSION . . . . . . . . . . . . . . . . . . . . . . . . 56 HYDROCHLORIC ACID (LOW PH — STRONG ACID) CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 NAPHTHENIC ACID CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 VI. NAPHTHENIC ACID CORROSION CHARACTERISTICS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 NAPHTHENIC ACID CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 VII. HIGH PH CORROSION CHARACTERISTICS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 VIII. MISCELLANEOUS EXAMPLES OF CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 MISCELLANEOUS EXAMPLES OF CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 EROSION — CORROSION IN A DEA STRIPPER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 CRUDE STORAGE TANK — CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 CRUDE PREHEAT CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 PRODUCT PIPELINE CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 ALUMINUM BRASS STRESS CRACKING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 UAN CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 FUEL STORAGE TANK CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 IX. WATERSIDE FAILURES IN THE REFINING INDUSTRY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .81 CASE HISTORY 1: FAILURE OF A STEAM GENERATOR TUBE DUE TO CAUSTIC CORROSION . . . . . . . . . . .81 CASE HISTORY 2: IDLE TIME OXYGEN CORROSION IN A POWER BOILER . . . . . . . . . . . . . . . . . . . . . . . . . .84 CASE HISTORY 3: CORROSION FATIGUE CRACKING OF A STEAM GENERATOR TUBE . . . . . . . . . . . . . . . .87 CASE HISTORY 4: FAILURE OF A BOILER DEAERATOR DUE TO STRESS CORROSION CRACKING . . . . . . .91 CASE HISTORY 5: OVERHEATING IN A POWER BOILER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .93 CASE HISTORY 6: FAILURE OF A CO BOILER ECONOMIZER TUBE DUE TO FIRESIDE CORROSION . . . . . .97 CASE HISTORY 7: FAILURE OF A HEAT EXCHANGER TUBE DUE TO STRESS CORROSION CRACKING . .101 CASE HISTORY 8: STRESS CORROSION CRACKING OF A BRASS HEAT EXCHANGER TUBE . . . . . . . . . . .103 CASE HISTORY 9: FAILURE OF A HEAT EXCHANGER TUBE DUE TO DEALLOYING AND CORROSION FATIGUE CRACKING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .107 CASE HISTORY 10: FAILURE OF A HEAT EXCHANGER TUBE DUE TO MICROBIOLOGICALLY INFLUENCED CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .109 CASE HISTORY 11: FAILURE OF A HEAT EXCHANGER TUBE DUE TO CONCENTRATION CELL CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .112 ABOUT THE AUTHORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

i

LIST OF ILLUSTRATIONS ILLUSTRATIONS FIGURE I-1: UNIFORM CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 FIGURE I-2: UNIFORM CORROSION IN TUBING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 FIGURE I-3: UNIFORM CORROSION WITH NUMEROUS PITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 FIGURE I-4: UNIFORM CORROSION SHOWING HILLS AND VALLEYS . . . . . . . . . . . . . . . . . . . . . . . . 3 FIGURE I-5: UNIFORM CORROSION SHOWING HILLS AND VALLEYS . . . . . . . . . . . . . . . . . . . . . . . . 4 FIGURE I-6: CO2 ATTACK — HONEYCOMB PATTERN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 FIGURE I-7: CO2 ATTACK — HONEYCOMB PATTERN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 FIGURE I-8: HONEYCOMB ATTACK IN A CHOKE BODY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 FIGURE I-9: CO2 PITTING CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 FIGURE I-10:TUBING PERFORATED BY CO2 CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 FIGURE I-11:TUBING PERFORATED BY CO2 CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 FIGURE I-12:WORMHOLE ATTACK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 FIGURE I-13: SUCKER ROD CONNECTOR WITH WORMHOLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 FIGURE I-14:TUBING WITH WORMHOLE ATTACK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 FIGURE I-15: CLOSE-UP TUBING WITH WORMHOLE ATTACK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 FIGURE I-16:WORMHOLE EFFECT — CLUSTER OF LARGE PITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 FIGURE I-17: RINGWORM CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 FIGURE I-18: SUSPECTED RINGWORM CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 FIGURE I-19: PARTED TUBING BY RINGWORM ATTACK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 FIGURE I-20: SEVERE THINNING CAUSED BY RINGWORM ATTACK . . . . . . . . . . . . . . . . . . . . . . . . 11 FIGURE I-21: HEAT-AFFECTED ZONE CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 FIGURE I-22: HAZ AS TYPIFIED BY DEEP CORROSION ALONG WELD SEAM . . . . . . . . . . . . . . . . . 12 FIGURE I-23: CLOSE-UP OF FIGURE I-22 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 FIGURE I-24: PHOTOMICROGRAPH OF HAZ ATTACK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 FIGURE I-25: CORRODED WELD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 FIGURE I-26: CORROSION RESULTING FROM ROLLED-IN INCLUSIONS . . . . . . . . . . . . . . . . . . . . . 14 FIGURE I-27: CORROSION RESULTING FROM ROLLED-IN INCLUSIONS . . . . . . . . . . . . . . . . . . . . . 15 FIGURE I-28: MESA ATTACK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 FIGURE I-29: MESA ATTACK ON TUBING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 FIGURE I-30: MESA ATTACK — NOTE DEEP GOUGES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 FIGURE I-31: RAINDROP ATTACK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 FIGURE I-32: RAINDROP ATTACK — DEEP PITS WITH TAILS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 FIGURE I-33: CLOSE-UP VIEW OF PIT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 FIGURE I-34: EXAMPLE OF THE RAINDROP EFFECT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 FIGURE I-35: EXAMPLE OF THE RAINDROP EFFECT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 FIGURE I-36: EXAMPLE OF THE RAINDROP EFFECT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 FIGURE I-37: FLOW-ENHANCED CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 FIGURE I-38: FLOW-ENHANCED CORROSION IN AN ELBOW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 FIGURE I-39: CLOSE-UP SHOWING SEVERE METAL LOSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 FIGURE I-40: CLOSE-UP SHOWING PENETRATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 FIGURE I-41: COMBINATION OF CORROSION MECHANISMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 FIGURE I-42: SEVERE METAL LOSS IN BEND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 FIGURE I-43: FLOW-ENHANCED CORROSION IN U-BEND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 FIGURE I-44: FLOW-ENHANCED CORROSION AT NIPPLE JOINTS . . . . . . . . . . . . . . . . . . . . . . . . . . 23 FIGURE I-45: FLOW-ENHANCED CORROSION AT NIPPLE JOINTS . . . . . . . . . . . . . . . . . . . . . . . . . . 24

ii

LIST OF ILLUSTRATIONS (CONTINUED) FIGURE I-46: SAND EROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 FIGURE I-47:WIRELINE ATTACK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 FIGURE I-48:WIRELINE ATTACK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 FIGURE I-49: A PIT FORMED FROM A PORE IN THE COATING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 FIGURE I-50: CREVICE CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 FIGURE I-51: CREVICE CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 FIGURE I-52: CORROSION FATIGUE IN DRILL PIPE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 FIGURE I-53: CORROSION FATIGUE IN SUCKER RODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 CHART 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 FIGURE II-1: OXYGEN CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 FIGURE II-2: OXYGEN CORROSION IN A FLOWLINE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 FIGURE II-3:TYPICAL PITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 FIGURE II-4: SHALLOW SAUCER-LIKE PITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 FIGURE II-5: AREAS OF METAL LOSS ON SUCKER RODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 FIGURE II-6: AREAS OF METAL LOSS ON SUCKER RODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 FIGURE II-7:TYPICAL RUST-COLORED DEPOSITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 FIGURE II-8: SAUCER-LIKE DEPOSITS AND PITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 FIGURE II-9: OXYGEN CORROSION IN A 36" PIPELINE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 CHART 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 FIGURE III-1: HYDROGEN SULFIDE (H2S) CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 FIGURE III-2: EXAMPLE OF H2S ATTACK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 FIGURE III-3: FES CRYSTALS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 FIGURE III-4: H2S ATTACK ON SUCKER RODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 FIGURE III-5: H2S ATTACK ON SUCKER RODS FOLLOWED BY CORROSION FATIGUE BREAK . . 38 FIGURE III-6: SULFIDE STRESS CRACKING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 FIGURE III-7: CORROSION FATIGUE IN SUCKER ROD CONNECTORS . . . . . . . . . . . . . . . . . . . . . . 39 FIGURE III-8: CRACK PROPAGATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 FIGURE III-9: STRESS CRACKS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 FIGURE III-10: A FAILURE CAUSED BY SULFIDE STRESS CRACKING OF A KILL STRING . . . . . . . . . 41 FIGURE III-11: SULFIDE STRESS CRACKING OF A CASING COLLAR . . . . . . . . . . . . . . . . . . . . . . . . 42 CHART 3: SULFIDE STRESS CRACKING REGION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 FIGURE III-12: HYDROGEN EMBRITTLEMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 FIGURE III-13: CHLORIDE STRESS CORROSION IN STAINLESS STEEL . . . . . . . . . . . . . . . . . . . . . . . 44 CHART 4: APPROACHES TO STRESS CORROSION CONTROL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 FIGURE IV-1: MICROORGANISM INFLUENCED CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 FIGURE IV-2: BACTERIAL ATTACK CHARACTERISTICS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 FIGURE IV-3: BACTERIAL ATTACK CHARACTERISTICS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 FIGURE IV-4:THE TERRACED EFFECT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 FIGURE IV-5: ANOTHER EXAMPLE OF THE TERRACED EFFECT . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 FIGURE IV-6: PIPE CLEANER TEST EXAMPLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 FIGURE IV-7: PIPE CLEANER TEST EXAMPLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 FIGURE IV-8: MIC ATTACK IN CONJUNCTION WITH CO2 ATTACK . . . . . . . . . . . . . . . . . . . . . . . . 51 FIGURE IV-9: MIC DETECTED IN THE 3 O'CLOCK POSITION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 FIGURE IV-10: MIC DETECTED IN THE 9 O'CLOCK POSITION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 FIGURE IV-11: DETECTABLE BLACK GROWTH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 FIGURE V-1: HYDROCHLORIC ACID AND HYDROGEN SULFIDE CORROSION . . . . . . . . . . . . . . . 57

iii

LIST OF ILLUSTRATIONS (CONTINUED) FIGURE V-2: CLOSER INSPECTION OF THE PITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 FIGURE V-3: 13 CR STAINLESS STEEL (UNS410) BUBBLE CAP CORROSION . . . . . . . . . . . . . . . . . . . 58 FIGURE V-4: CLOSE-UP OF PITTING CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 FIGURE V-5: AMMONIUM CHLORIDE CORROSION ON CARBON STEEL HEAT EXCHANGER TUBES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 FIGURE V-6: PITS WITH NARROW OPENINGS AND DEEP CAVERNOUS VOIDS . . . . . . . . . . . . . . . 59 FIGURE V-7: AMMONIUM CHLORIDE CORROSION ON SS410 TRAYS . . . . . . . . . . . . . . . . . . . . . . . 60 FIGURE V-8: AMMONIUM CHLORIDE CORROSION ON SS410 TRAYS . . . . . . . . . . . . . . . . . . . . . . . 60 FIGURE V-9: 304L STAINLESS STEEL TRAY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 FIGURE V-10: STRESS CORROSION CRACKING FAILURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 FIGURE V-11: CARBON STEEL TRAYS — HCI CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 FIGURE V-12: UNDERDEPOSIT CORROSION — NOTE THE THINNING CLEANED SURFACE . . . . 62 FIGURE V-13: SULFIDATION/OXIDATION OR AMMONIUM CHLORIDE . . . . . . . . . . . . . . . . . . . . . . 63 FIGURE V-14. UNDERDEPOSIT CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 FIGURE V-15: PITTING AND GOUGING — EXTERNAL SURFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 FIGURE VI-1: NAPHTHENIC ACID ATTACK ON 14CR TRAY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 FIGURE VI-2:TYPICAL PITTING FROM NAPHTHENIC ACID CORROSION ON HEATER PLUGS . . . 66 FIGURE VII-1: DE-ALLOYING TUBES — H2S CORROSION 90-10 CU:NI HEAT EXCHANGER TUBES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 FIGURE VII-2: CLOSE-UP OF DE-ALLOYING TUBES — H2S CORROSION 90-10 CU:NI HEAT EXCHANGER TUBES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 FIGURE VIII-1: EROSION CORROSION — CARBON STEEL TEE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 FIGURE VIII-2: CLOSE-UP PITS AFTER CLEANING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 FIGURE VIII-3: EROSION-CORROSION — CARBON STEEL OVERHEAD EXCHANGER TUBES . . . . 71 FIGURE VIII-4: EROSION-CORROSION — DEA STRIPPER TOWER WALL . . . . . . . . . . . . . . . . . . . . . 72 FIGURE VIII-5: LARGE AREAS ERODED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 FIGURE VIII-6: CRUDE STORAGE TANK FAILURE MOST LIKELY FROM WET OXYGENATED SOILS 73 FIGURE VIII-7: OXYGEN, H2S, AND CO2 CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 FIGURE VIII-8: CARBON STEEL PRODUCT PIPELINE CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . 75 FIGURE VIII-9: CLOSE-UP OF CARBON STEEL PRODUCT PIPELINE CORROSION . . . . . . . . . . . . . . 75 FIGURE VIII-10: ALUMINUM BRASS STRESS CRACKING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 FIGURE VIII-11: ALUMINUM BRASS STRESS CRACKING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 FIGURE VIII-12: UAN CORROSION — CARBON STEEL STORAGE TANK . . . . . . . . . . . . . . . . . . . . . 77 FIGURE VIII-13. CLOSE-UP VIEW OF THE CORROSION AT A PLATE BOUNDARY . . . . . . . . . . . . . . 77 FIGURE VIII-14: UAN CORROSION ON CARBON STEEL STORAGE TANKS . . . . . . . . . . . . . . . . . . . 78 FIGURE VIII-15: FUEL STORAGE TANK CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 FIGURE VIII-16: FUEL STORAGE TANK CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 FIGURE VIII-17: FUEL STORAGE TANK CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 FIGURE VIII-18: FUEL STORAGE TANK CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 FIGURE IX-1: LOCALIZED METAL LOSS AT END OF TUBE THAT CONTACTED BOTTOM TUBESHEET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .81 FIGURE IX-2:TRANSFER LINE EXCHANGER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .82 FIGURE IX-3:TUBESHEET WHERE CORROSION OCCURRED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .82 FIGURE IX-4: CORROSION PRODUCTS IN WASTED AREA CONTAINING MAGNETITE NEEDLES AND METALLIC COPPER PARTICLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .83

iv

LIST OF ILLUSTRATIONS (CONTINUED) FIGURE IX-5: BOTTOM-HALF SECTION SHOWING INTERNAL SURFACE COVERED WITH UNIFORM LAYER OF BLACK IRON OXIDE INTERRUPTED BY SCATTERED IRON OXIDE TUBERCLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .84 FIGURE IX-6: IRON OXIDE TUBERCLE WITH NO FLOW ORIENTATION . . . . . . . . . . . . . . . . . . . . .85 FIGURE IX-7: DEPRESSIONS FORMED AT TUBERCLE SITES REVEALED BY CLEANING . . . . . . . . . . .85 FIGURE IX-8: LOCALIZED STRIP OF ORANGE-BROWN IRON OXIDE ALONG BOTTOM OF HORIZONTAL TUBE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .86 FIGURE IX-9: HEAVY DEPOSIT LAYERS COVERING INTERNAL SURFACE OF BOILER TUBE . . . . . .87 FIGURE IX-10: NUMEROUS TRANSVERSELY ORIENTED FISSURES ON INTERNAL SURFACE . . . . . .88 FIGURE IX-11: FISSURE CAUSED BY CORROSION FATIGUE CRACKING FILLED WITH DENSE OXIDE CONTAINING CENTERLINE CRACK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .88 FIGURE IX-12: ILLUSTRATION OF THE MECHANISM OF CORROSION FATIGUE CRACKING . . . . .89 FIGURE IX-13: NUMEROUS FISSURES INITIATING FROM SHALLOW DEPRESSIONS CAUSED BY OLD IDLE TIME OXYGEN CORROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .90 FIGURE IX-14: CRACKS ON INTERNAL SURFACE COVERED WITH IRON OXIDE CORROSION PRODUCTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .91 FIGURE IX-15: IRON OXIDE CORROSION PRODUCTS REMOVED TO REVEAL DEPRESSIONS . . . .92 FIGURE IX-16: METALLOGRAPHIC CROSS-SECTION REVEALING BRANCHED,TRANSGRANULAR CRACKS PENETRATING OUTWARD FROM INTERNAL SURFACE . . . . . . . . . . . . . . .92 FIGURE IX-17: LOCALIZED BULGES ON HOT SIDE OF WALL TUBE . . . . . . . . . . . . . . . . . . . . . . . . .94 FIGURE IX-18: LOCALIZED DEPOSITION AT BULGE SITE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .94 FIGURE IX-19: MICROSTRUCTURE OF TUBE IN AS-MANUFACTURED CONDITION IN ALL LOCATIONS EXCEPT AT BULGES CONSISTING OF COLONIES OF PEARLITE IN FERRITE MATRIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .95 FIGURE IX-20: MICROSTRUCTURE AT BULGES CONSISTING OF SPHEROIDAL CARBIDES IN FERRITE MATRIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .95 FIGURE IX-21: INTERGRANULAR CREEP FISSURES ON EXTERNAL SURFACE AT BULGE . . . . . . . .96 FIGURE IX-22: METAL STRENGTH VERSUS TEMPERATURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .96 FIGURE IX-23: EXTERNAL SURFACE DEPRESSIONS AND ACIDIC, LIGHT-COLORED DEPOSITS . . .98 FIGURE IX-24: DEPOSITS REMOVED TO REVEAL DEPRESSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . .98 FIGURE IX-25: SULFURIC ACID DEWPOINT AS A FUNCTION OF FUEL SULFUR CONTENT . . . . .99 FIGURE IX-26: SULFURIC ACID DEWPOINT AS A FUNCTION OF WATER VAPOR CONCENTRATION IN FURNACE ENVIRONMENT . . . . . . . . . . . . . . . . . . . . . . . . . . .99 FIGURE IX-27:WALL OF ADMIRALTY BRASS HEAT EXCHANGER TUBE PENETRATED BY A CRACK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .101 FIGURE IX-28: BRANCHED, INTERGRANULAR CRACK AND FISSURES PENETRATING FROM INTERNAL SURFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .102 FIGURE IX-29: BRITTLE FRACTURE EDGE CONTAINING MULTIPLE CRACK INITIATION SITES . .103 FIGURE IX-30: NUMEROUS TRANSVERSELY ORIENTED FISSURES NEAR FRACTURE EDGE . . . . . .104 FIGURE IX-31:TIGHT BRANCHED FISSURE CAUSED BY STRESS CORROSION CRACKING . . . . .105 FIGURE IX-32:TRANSGRANULAR PROPAGATION PATH OF A FISSURE . . . . . . . . . . . . . . . . . . . . .105 FIGURE IX-33: INCREASE IN TIME TO FAILURE WITH DECREASING STRESS AND THRESHOLD STRESS BELOW WHICH CRACKING DOES NOT OCCUR . . . . . . . . . . . . . . . . . . . .106 FIGURE IX-34: CRACKING FAILURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .107 FIGURE IX-35: INTERNAL SURFACE DEPRESSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .108 FIGURE IX-36: METALLIC COPPER COVERING SURFACE, INDICATING DEZINCIFICATION . . . . .108 FIGURE IX-37: CORROSION FATIGUE FISSURES PENETRATING FROM BASE OF DEPRESSION . . .109

v

LIST OF ILLUSTRATIONS (CONTINUED) FIGURE IX-38: SCATTERED, DEEP EXTERNAL SURFACE DEPRESSIONS . . . . . . . . . . . . . . . . . . . . . .110 FIGURE IX-39: CLUSTERS OF MUTUALLY-INTERSECTING, ROUNDED DEPRESSIONS . . . . . . . . . .110 FIGURE IX-40: METALLOGRAPHIC CROSS-SECTION OF DEPRESSION COVERED WITH CORROSION PRODUCTS CONTAINING SULFIDE . . . . . . . . . . . . . . . . . . . . . . . . . .111 FIGURE IX-41: IRON OXIDE LAYERS AND TUBERCLES OVERLAID WITH WATER-FORMED DEPOSIT LAYERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .112 FIGURE IX-42: DEPOSIT AND CORROSION PRODUCT LAYERS REMOVED TO REVEAL ROUGH-SURFACED, UNDERCUT DEPRESSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . .113 FIGURE IX-43: METALLOGRAPHIC CROSS-SECTION OF ROUGH-SURFACED, UNDERCUT DEPRESSIONS BENEATH HOLLOW IRON OXIDE TUBERCLE . . . . . . . . . . . . . . . . . .113 FIGURE IX-44: DIFFERENTIAL OXYGEN CONCENTRATION CELL . . . . . . . . . . . . . . . . . . . . . . . . .114 FIGURE IX-45: ION CONCENTRATION CELL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .115 FIGURE IX-46:TUBERCLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .116

vi

I. CARBON DIOXIDE CORROSION CHARACTERISTICS “SWEET CORROSION” Form of Corrosion Uniform — FeCO3 film Pits — Round to oval shape Sharp edges Round bottoms Smooth sides and bottoms Corrosion Products Iron carbonate (FeCO3), siderite Iron oxide (Fe3O4, magnetite Color — tan, green, brown to black Corrosion Reactions CO2 + H2O H2CO3 + Fe 3Fe + 4H2O






H2CO3 FeCO3 + 2H+ + 2eFe3O4 + 8H+ + 8e-

Unique Forms of Corrosion Wormhole attack Ringworm corrosion Heat-affected corrosion Mesa attack Rain drop Treatment Drilling — pH control with caustic soda Producing wells — Corrosion inhibitors Flowlines — Continuous corrosion inhibitor injection

1

CARBON DIOXIDE CORROSION Carbon dioxide (CO2) is absorbed in brines and can cause severe corrosive attack of steel.The corrosion rates of CO2 can theoretically be from 1,000 to 10,000 mils per year, if not stopped by corrosion products or inhibitors. Uniform corrosion takes the form of hills and valleys with very little area that is not attacked (Figure I-1).

Figure I-1: Uniform Corrosion

An example of uniform CO2 corrosion from tubing is shown in Figure I-2.

Figure I-2: Uniform Corrosion in Tubing

2

Another form of CO2 corrosion takes the form of numerous pits (Figure I-3).

Figure I-3: Uniform Corrosion with Numerous Pits

Figure I-4 and Figure I-5 show additional examples of uniform corrosion, leaving a distinctive hill and valley texture.

Figure I-4: Uniform Corrosion Showing Hills and Valleys

3

Figure I-5: Uniform Corrosion Showing Hills and Valleys

Another form of CO2 attack has been observed on exposed faces of high-pressure wells, taking the form of an intricate honeycomb (Figure I-6 and Figure I-7).

Figure I-6: CO2 Attack — Honeycomb Pattern

4

Figure I-7: CO2 Attack — Honeycomb Pattern

Another example of honeycomb attack can be seen on a choke body (Figure I-8).

Figure I-8: Honeycomb Attack in a Choke Body

5

CO2 dissolved in brine forms carbonic acid, and carbonic acid is extremely corrosive to steel, with horrendous corrosion rates. However, the carbonic acid corrosion by-product, iron carbonate (FeCO3), usually forms protective films. FeCO3 is relatively insoluble at downhole operating conditions. However, at breaks in the FeCO3 film pits do develop.This pitting corrosion is extremely rapid and can cause failures. CO2 pitting corrosion can be seen in Figure I-9.

Figure I-9: CO2 Pitting Corrosion

A piece of downhole tubing that perforated within two years of operation is shown in Figure I-10 and Figure I-11.

Figure I-10: Tubing Perforated by CO2 Corrosion

6

Figure I-11: Tubing Perforated by CO2 Corrosion

Another typical formation for CO2 corrosion is called wormhole type of attack (Figure I-12).

Figure I-12: Wormhole Attack

7

In the wormhole attack, individual CO2 pits are interconnected and form a wormhole appearance, with segments very similar to the segmentation pattern on a worm. A sucker rod connector had a typical wormhole (Figure I-13).

Figure I-13: Sucker Rod Connector with Wormhole

Tubing with wormhole attack can be seen in Figure I-14, with a close-up view in Figure I-15.

Figure I-14: Tubing with Wormhole Attack

8

Figure I-15: Close-up Tubing with Wormhole Attack

At times, the wormhole can take the form of a cluster of large pits (Figure I-16).

Figure I-16: Wormhole Effect — Cluster of Large Pits

9

Another form of corrosion observed in the oilpatch is named ringworm corrosion (Figure I-17).

Figure I-17: Ringworm Corrosion

Ringworm corrosion is a metallurgical effect caused by the making of the upset. In fabricating the upset, a zone is created 4 to 6 inches from the upset shoulder, wherein carbon particles have been spheroidized.These spheroidalshaped particles accelerate corrosion in mildly corrosive or even non-corrosive environments.This is a galvanic effect. If one observes numerous through-the-wall penetrations in tubing 4 to 6 inches from the upset shoulders such as in Figure I-18, ringworm corrosion is suspected.

Figure I-18: Suspected Ringworm Corrosion

10

Ringworm can cause tubing parting (Figure I-19).

Figure I-19: Parted Tubing by Ringworm Attack

Another example shows severe thinning at this point, with a smooth surface (Figure I-20).

Figure I-20: Severe Thinning Caused by Ringworm Attack The treatment for ringworm corrosion is to normalize the steel after making the upset. Normalizing redissolves the carbide particles eliminating the zone created by making the upset.

11

Another form of galvanically accelerated corrosion is called heat-affected zone corrosion (Figure I-21).

Figure I-21: Heat-affected Zone Corrosion

Heat-affected zone corrosion or HAZ attack is typified by deep corrosion along the weld seam. Figure I-22 shows a typical example and Figure I-23 illustrates a close-up view.

Figure I-22: HAZ as Typified by Deep Corrosion Along Weld Seam

12

Figure I-23: Close-up of Figure I-22

The weld seam is relatively unharmed, although the zone right next to the weld is deeply corroded.This can be seen in a photomicrograph through the weld and heat-affected zone.The heat-affected zone is darker than the weld (Figure I-24).

Figure I-24: Photomicrograph of HAZ Attack

13

The HAZ corrosion mechanism is similar to ringworm corrosion. During welding, a zone is created wherein carbon particles accumulate and precipitate to form nodules.These nodules in mildly corrosive environments form a galvanic cell.The weld itself can preferentially corrode at times (Figure I-25).

Figure I-25: Corroded Weld

Heat-affected zone corrosion is found in elbows, reducers, and flowline tees.These are typically field welds, where there are no means of negating the effect of welding by post weld heat treatments.The fix for heat-affected zone corrosion is to normalize the steel after welding, which can be done on tubing, but may not be possible on field welds.

Figure I-26: Corrosion Resulting from Rolled-in Inclusions

14

An unusual form of metallurgical attack on the outside of tubing is shown in Figure I-26 and I-27, and is attributed to rolled-in inclusions.

Figure I-27: Corrosion Resulting from Rolled-in Inclusions

Another form of CO2 corrosion is called mesa attack. Mesa attack usually occurs where areas of metal that have a good protective coating exist and where other areas have no protective coating. In addition, mesa attack occurs in flowing environments (Figure I-28).

Figure I-28: Mesa Attack

15

Tubing that has undergone mesa attack can be seen in Figure I-29. Flat surfaces and deep gouges between the flat surfaces typify it (Figure I-30). Mesa corrosion can be controlled with inhibitor films to get greater protection on those surfaces where corrosion has occurred.

Figure I-29: Mesa Attack on Tubing

Figure I-30: Mesa Attack — Note Deep Gouges

16

Another form of CO2 corrosion occurs in gas condensate wells, which do not produce a large amount of water. In these wells, a zone where water droplets, or water of condensation, will drop out.This phenomenon is labeled raindrop attack (Figure I-31).

Figure I-31: Raindrop Attack

In this area, very deep pits occur with tails.These deep pits on tubing are shown in Figure I-32.This tubing was the second and third joint below the Christmas tree.

Figure I-32: Raindrop Attack — Deep Pits with Tails

17

A close-up view of the pit shows the typical tail.The tail is usually downstream of the flow, pointing opposite to the direction of flow. It is formed by the escaping or dripping of corrosion by-products out of the deep pit (Figure I-33).

Figure I-33: Close-up View of Pit

Other examples of the raindrop effect are shown in Figure I-34, Figure I-35, and Figure I-36. Use of corrosion inhibitor films might prevent the raindrop effect by preventing droplets from hanging up in the tubing.The most prevalent place for the raindrop effect has been relatively close to the Christmas tree head. However, the effect has also occurred at lower levels where the temperature is low enough to condense the moisture from the gas stream.

Figure I-34: Example of the Raindrop Effect

18

Figure I-35: Example of the Raindrop Effect

Figure I-36: Example of the Raindrop Effect

19

Another form of severe metal loss usually leading to rapid failures is flow-enhanced corrosion (Figure I-37).

Figure I-37: Flow-Enhanced Corrosion Flow-enhanced corrosion was at one time called erosion-corrosion. Flow-enhanced corrosion usually occurs when a change of direction in the flow of fluids occurs, such as near an elbow.Typically, the mechanism follows a deposition of a protective FeCO3 film with rubbing off the film by the flow of fluids, and sometimes solids. After the protective coating is abraded off, a new surface is exposed which then immediately starts to corrode.The corrosion product is a softer surface, which is then eroded away, and the process continues until a failure occurs.The following illustrates an elbow taken from a flowline (Figure I-38).

Figure I-38: Flow-Enhanced Corrosion in an Elbow

20

Figure I-39, a close-up view after cleaning, shows the severe metal loss. Figure I-40 shows the through-the-wall penetration.

Figure I-39: Close-up Showing Severe Metal Loss

Figure I-40: Close-up Showing Penetration

21

In Figure I-41, this flowline has an unusual failure mode. It is a combination of flow-enhanced corrosion and heat-affected zone attack.

Figure I-41: Combination of Corrosion Mechanisms

After a 45º bend, severe metal loss was noted in the bend (Figure I-42). In addition, a clear demarcation can be found where the flow had ceased, and there was no further attack on the pipe.

Figure I-42: Severe Metal Loss in Bend

22

Another example of this flow-enhanced corrosion is located at the top of a U-bend, as seen in this flowline (Figure I-43).

Figure I-43: Flow-Enhanced Corrosion in U-bend

Another form of flow-enhanced corrosion is at nipple joints. Usually there is a gap when a threaded joint is made up, and at the gap, turbulence is generated and flow-enhanced corrosion can occur (Figure I-44 and Figure I-45).

Figure I-44: Flow-Enhanced Corrosion at Nipple Joints

23

Figure I-45: Flow-Enhanced Corrosion at Nipple Joints

Sand erosion can also occur (Figure I-46).This erosion is typified by a smooth surface with a sand dune pattern.

Figure I-46: Sand Erosion

24

On coated tubing, wireline operations can sometimes cut the protective coating, resulting in corrosive attack where the breaks occur (Figure I-47 and Figure I-48).This attack is termed wireline attack.

Figure I-47: Wireline Attack

Figure I-48: Wireline Attack

25

Pores can sometimes cause an undercutting and pits in the base metal (Figure I-49).

Figure I-49: A Pit Formed from a Pore in the Coating

Another form of attack is called crevice corrosion. Crevice corrosion usually occurs in metals with a good protective coating. However, at crevices, there are places where the protective coating does not get underneath, and severe localized corrosion can occur (Figure I-50 and Figure I-51).

Figure I-50: Crevice Corrosion

26

Figure I-51: Crevice Corrosion

Corrosion fatigue is a combination of corrosion and cyclic stresses (Figure I-52).

Figure I-52: Corrosion Fatigue in Drill Pipe

27

In the oilfield, the two places where corrosion fatigue is prevalent is in drill pipe and sucker rods. Both of these places are under severe cyclic stresses (Figure I-50 and Figure I-51).

Figure I-53: Corrosion Fatigue in Sucker Rods

This failure has all the attributes of a corrosion fatigue failure, with an initiation point, a smooth half moon thumbnail, a crystalline area, and a lip break (Figure I-53).

The arrowed points indicate the unbroken specimens. 110 o o 100 o Normal air. x x Natural sea water. x 90 80 70 x x o 60 50 o x x o x ox o Tmin o o xo 40 o o x x oo xx oxo o o oo 30 o o x x x xx x x o 20 x x xo 10 0 10 4 10 5 10 6 10 7 Stress amplitude

Stress Amplitude om=o, ton/in 2

Chart 1 demonstrates the relationship between grades to failure and applied stress.With corrosion, a safe operating stress does not exist.

NUMBER OF CYCLES (N)

Chart 1

28

m Time

x xx x

II. OXYGEN CORROSION CHARACTERISTICS Forms of Attack Uniform — With red rust deposit Pits — Smooth bottoms Smooth sides Sloped edges Width much greater than pit depth Corrosion Products Iron oxides

FeO(OH), geothite Fe2O3, hematite Fe3O4, magnetite FeO(OH), ferrous hydroxide

Corrosion Reactions 2Fe + 3H2O 3Fe + 4H2O Fe + 2H2O






Fe2O3 + 6H+ Fe3O4 + 8H+ FeO(OH) + 3H+

Unique Forms of Corrosion Oxygen concentration cells Tubercules Treatment Drilling — Oxygen scavengers Producing wells — Corrosion inhibitors; elimination of oxygen sources, oxygen scavengers Flowlines — Corrosion inhibitors; elimination of oxygen sources, oxygen scavengers

29

OXYGEN CORROSION Oxygen is an extremely corrosive gas in produced brines, even more corrosive than CO2. Oxygen corrosion (Figure II1) usually results in iron oxide corrosion by-products, usually hematite, goethite, and magnetite.

Figure II-1: Oxygen Corrosion

Typically, the corrosive by-products are a reddish brown rust (Figure II-2).

Figure II-2: Oxygen Corrosion in a Flowline

30

A close-up view shows the typical pits (Figure II-3).

Figure II-3: Typical Pits

On cleaning, the shallow saucer-like pits are seen (Figure II-4).

Figure II-4: Shallow Saucer-like Pits

31

Oxygen corrosion on sucker rods are large areas of metal loss (Figure II-5 and Figure II-6).

Figure II-5: Areas of Metal Loss on Sucker Rods

Figure II-6: Areas of Metal Loss on Sucker Rods

32

Typical rust-colored deposits taken from a water line in a mining operation are shown in Figure II-7.

Figure II-7: Typical Rust-colored Deposits

After cleaning, saucer-like deposits are seen (Figure II-8).

Figure II-8: Saucer-like Deposits and Pits

33

Another example of oxygen corrosion is in this 36" pipeline (Figure II-9).

Figure II-9: Oxygen Corrosion in a 36" Pipeline

Oxygen is the most corrosive gas in the presence of water (see Chart 2). Although the maximum dissolved oxygen in the water is only 8 ppm, dissolved CO2 can increase to 800 ppm and dissolved H2S to 400 ppm. Preventing the ingress of oxygen into the system can usually control oxygen corrosion. In addition, use of oxygen scavengers are notably effective in drilling operations where oxygen is a significant problem. Oxygen scavengers are also very effective in salt-water disposal systems where oxygen ingress can cause serious downhole and flowline corrosion failures.

OVERALL CORROSION RATE OF CARBON STEEL MPY

25

25 C=76 F 5-7 Days 2-5 g/l NaCl <50 ppm HCO3

20

O2

CO 2

15

10

H2 S

5

O2

1

2

3

4

5

6

7

8

CO 2

100

200

300

400

500

600

700

800

H2 S

50

100

150

200

250

300

350

400

DISSOLVED GAS CONCENTRATION IN WATER PHASE, ppm

Chart 2

34

III. HYDROGEN SULFIDE CORROSION CHARACTERISTICS Forms of Attack Uniform — Pits —

FeS2 film Conical shaped Pit bottoms sharp Edges etched and slightly sloped Sides etched

Corrosion Products Black, blue-black iron sulfide Fe 1-x S pyrite, greigite, mackinwaite Kansite, iron oxide (Fe3O4), magnetite, sulfur (S) Sulfur dioxide (SO2) Corrosion Reactions H2S(g) + Fe + H2O
Fe 1-x S (Prec.) + 2H+ Fe 1-x S + O2
Fe3O4 + SO2 Unique Forms of Corrosion Sulfide stress cracking Hydrogen blistering Hydrogen embrittlement Stepwise cracking Treatment Drilling — High pH, Zn treatments Production — Corrosion inhibitors Flowlines — Corrosion inhibitors, hydrogen sulfide scavengers

35

HYDROGEN SULFIDE CORROSION Hydrogen sulfide (H2S) in produced brines generates hydrogen ions and HS ions.The hydrogen ions, being acidic, attack steel. H2S corrosion is typified by pitting, which takes the form of shallow round depressions with etched bottoms (Figure III-1).

Figure III-1: Hydrogen Sulfide (H2S) Corrosion In addition, the sides of the pits are sloped. A quick test to detect H2S corrosion is to test the fluids for H2S.A few drops of hydrochloric acid with arsenious oxide on the surface deposits turns yellow if FeS2 is present. Some examples of the hydrogen sulfide attack are shown in Figure III-2.

Figure III-2: Example of H2S Attack 36

Also associated with H2S attack are some crystals of FeS2 (Figure III-3).

Figure III-3: FeS Crystals

37

H2S attack on sucker rods can cause breaks because of corrosion fatigue (Figure III-4 and Figure III-5).

Figure III-4: H2S Attack on Sucker Rods

Figure III-5: H2S Attack on Sucker Rods Followed by Corrosion Fatigue Break

38

Sulfide stress cracking occurs when H2S corrosion is accelerated by stresses (Figure III-6).

Figure III-6: Sulfide Stress Cracking

Alternating stresses can propagate notches, which are initiated by corrosion. An example of corrosion fatigue occurred in sucker rod connectors (Figure III-7).

Figure III-7: Corrosion Fatigue in Sucker Rod Connectors

39

A cross-section shows how deep the cracking can propagate into the metal, with a very small opening (Figure III-8 and Figure III-9).

Figure III-8: Crack Propagation

Figure III-9: Stress Cracks

40

An unusual instance of sulfide stress cracking occurred with H2S in a kill string. During acidizing, the ends of the kill string cracked where the connectors were hardened.The field used brine, which contained H2S, with the acid.The acid inhibitor was not effective, and all the joints cracked (Figure III-10).

Figure III-10: A Failure Caused by Sulfide Stress Cracking of a Kill String

41

Drilling collars and the drilling operation are also subject to sulfide stress cracking. Drilling stresses can cause failures. Figure III-11 shows a failure of a casing collar attributed to H2S stress cracking.

Figure III-11: Sulfide Stress Cracking of a Casing Collar

NACE document MR 01-75 defines the sulfide stress cracking region based on H2S concentration and total line pressures (Chart 3). GRAINS H 2 S PER 100 SCF

10

1

10,000

100

1000

5000 4000 3000

0. 05

2000

A

I PS L IA RT PA

1000

A

ES

PR E

R

SU

TOTAL PRESSURE, PSIA

SULFIDE STRESS CRACKING REGION

B

100

65 PSIA TOTAL PRESSURE

10 .0001 1

.001

.01

10

100

MOL % H 2S IN GAS PPM H 2S IN GAS

.1

1

10

1000

10,000

100,000

Sour gas systems (see Paragraph 1.3.1).

Chart 3: Sulfide Stress Cracking Region

42

Hydrogen, either from the corrosion process or during manufacture, can enter the metal and concentrate internally in high-stress areas, making the steel extremely brittle. Hydrogen embrittlement fractures are catastrophic because very little metal loss occurs before the break. Figure III-12 is an example of hydrogen embrittlement of a drill collar. On high strength steels in acid environments a phenomenon called hydrogen embrittlement can occur. The mechanism is similar to sulfide stress cracking except H2S may not be present. Another occurrence of failures with high strength steel is hydrogen induced cracking (HIC), also known as stepwise cracking. In low strength steels, HIC can occur in wet, sour environments. Figure III-12 shows a cross-section of a pressure vessel plate with cracking initiated at an inclusion site. A related type of failure is stress-oriented hydrogen induced cracking (SOHIC). SOHIC occurs in tanks around welded outlets in corrosive service when the welds are highly stressed.

Figure III-12: Hydrogen Embrittlement

43

Austenic stainless steels such as SS 304 and SS 316 are susceptible to chloride stress cracking in hot aqueous chloridecontaining environments. An SS 304 flowline handling a chloride-containing brine cracked at the highly stressed threaded areas (Figure III-13).

Figure III-13: Chloride Stress Corrosion in Stainless Steel

44

Stress corrosion cracking has three main avenues of control, namely mechanical, metallurgical, and environmental (Chart 4).

STRESS CORROSION CRACKING CONTROL MECHANICAL

METALLURICAL

ENVIRONMENTAL

Avoid stress concentrators

Change alloy composition

Modify environment

Relieve fabrication stresses

Change alloy structure

Apply anodic or cathodic protection

Introduce surface compressive stresses

Use metallic or conversation coating

Add inhibitor

Reduce operating stresses

Use organic coating

Non-destructive testing implications for design

Modify temperature

Chart 4: Approaches to Stress Corrosion Control

45

IV. MICROORGANISM INFLUENCED CORROSION CHARACTERISTICS Forms of Attack Pits — Irregular shaped Sloped edges Satiny bottom Etched sides Terraced sides Corrosion Products Iron sulfides, Sulfate reducing bacteria (SRB) Slime Growths of bacteria Corrosion Mechanisms Cathodic depolarization Cathodic FeS2 Concentration cells Sour corrosion Unique Forms of MIC Stainless steel pitting Slime Plugging Treatment Drilling — Biocides Production — Biocides, chlorine dioxide Flowlines — Biocides, chlorine dioxide Cost considerations — Continuous vs. batch EPA Biostat vs. biocide

46

MICROORGANISM INFLUENCED CORROSION Another facet of corrosion is microorganism influenced corrosion (MIC) (Figure IV-1).

Figure IV-1: Microorganism Influenced Corrosion

47

MIC occurs under bacterial growths and the exact mechanism is still open to discussion. In some cases, the bacteria produce acids, which cause the corrosion. In other cases with SRB, H2S is produced, which can cause H2S corrosion. With CO2, some typical CO2 attacking can occur because of the breakdown of the film by the bacteria. Bacterial attack is usually characterized by rounded pits with etched sides, edges, and bottoms (Figure IV-2 and Figure IV-3).

Figure IV-2: Bacterial Attack Characteristics

Figure IV-3: Bacterial Attack Characteristics

48

At times, the pits have a terraced effect (Figure IV-4 and Figure IV-5).

Figure IV-4: The Terraced Effect

Figure IV-5: Another Example of the Terraced Effect

49

One of the quick tests that can be run for SRB is the pipe cleaner test. Examples with positive results for the pipe cleaner test are shown in Figure IV-6 and Figure IV-7.

Figure IV-6: Pipe Cleaner Test Example

Figure IV-7: Pipe Cleaner Test Examples

50

Other examples of MIC attack in conjunction with CO2 are shown in Figure IV-8.

Figure IV-8: MIC Attack in Conjunction with CO2 Attack

51

In most cases, MIC attack occurs at the bottom of the line where water accumulates. However, MIC has also been detected in the 3 o'clock to 9 o'clock position, where there was a possible water and oil interface (Figure IV-9 and Figure IV-10).

Figure IV-9: MIC Detected in the 3 o'clock Position

Figure IV-10: MIC Detected in the 9 o'clock Position

52

One of the failures had detectable black growths, which tested positive for SRB, and had corrosion pits beneath the growth (Figure IV-11).

Figure IV-11: Detectable Black Growth

53

OBTAINING AND ENUMERATING A SAMPLE OF SESSILE BACTERIA 1. Obtain the surface containing sessile bacteria and carry out the following procedure as quickly as possible (Note 1). 2. Using a scalpel that has been sterilized by soaking in ethyl or isopropyl alcohol, thoroughly scrape the surface (Note 2). 3. Place the scraping into a half-ounce bottle that contains 1 gm of glass beads (diameter = 2 mm) and 10 ml of sterilized field water or other appropriate fluid (Note 3) and is equipped with a tight screw cap (Note 4). 4.Vigorously shake the contents 500 times. 5. Using 1-ml volumes out of the half-ounce bottle, make serial dilutions into the appropriate media as described in your Nalco Microbiology Guide. 6. Report results in counts/cm2. Notes: 1. Speed is necessary to minimize exposure of the sample to air. Using many different surfaces is possible including pipe walls, pits on a pipe wall, studs from Robbins devices, and corrosion coupons. 2. Minimize the handling of the sample as much as possible. Samples should be held with pre-sterilized forceps, if possible. Scrape 0.5 to 2.0 cm2 of the surface. 3. If sterilized field water is not available, or is impractical, a sterilized solution of similar salinity with pH of 7.0 can be substituted. Following are the ingredients for this formulation: K2HPO4 KH2PO4 Cysteine HCl Tween 80 NaCl Water

6.06 g 0.4 g 0.5 g 1.0 ml 30 gm or 80 gm (1) Add up to 1,000 ml by volume

(1) Use the strength that more closely approximates the total brine strength of the field. Sterilized sealed half-ounce bottles containing this formulation and glass beads can be obtained from several commercial sources. 4. If the surface to be scraped is small enough (such as a stud formulation from a Robbins device or biofilm probe), the scraped area may be put into the half-ounce bottle after scraping. 5. By measuring the amount of surface area that was scraped and observing the number of vials in the serial dilution that indicate growth, the number of bacteria per unit of surface area can be determined.

54

V. HYDROCHLORIC ACID (STRONG ACID) CORROSION CHARACTERISTICS Corrosion Examples from Refining Industry Each example of hydrochloric acid (low pH — strong acid) corrosion can actually represent several different mechanisms contributing to the total corrosion attack.The following represents general examples from the refining industry. Form of Corrosion Uniform acid etching Localized — Round bottom pits caused by partial water wetting Under deposit — Salts such as ammonium chloride Corrosion Products Black iron sulfide, (FeS) Iron chloride (Fe Cl2) Corrosion Reactions 2H2O + 2HCl + Fe 
FeCl2 + 2H2O + H2 FeCl2 + H2S 
FeS + 2HCl Unique Forms of Corrosion Often referred to as orange peel look Chloride stress corrosion of stainless steel Treatment Reduce chlorides — Desalting, caustic Neutralize acid at dew point Dilute acid and dissolve salts with wash water Examples of Corrosion Locations Crude units Vacuum units Straight-run distillation Hydro-desulfurization (HDS) Alkylation NGL gas plants

55

HYDROCHLORIC ACID CORROSION — LOW pH AQUEOUS CORROSION Acid attack is the largest corrosion challenge on a crude distillation unit. Most of the corrosion damage to a crude unit occurs in the first one or two exchanger sets in the crude unit overhead. Some crude units are also challenged by corrosion in the top of the crude tower or even in the overhead piping before the first exchangers.The corrosion is always due to low pH acids, but the physical and chemical mechanism can vary. Several descriptions describe the mechanisms, but these expressions are frequently misused. This misuse has caused confusion in the industry, making it difficult to understand critical aspects of any corrosion control program. Following are the physical and chemical conditions that match these descriptions: • Shock Condensation Corrosion — This corrosion is associated with strong acid, low pH corrosion in an area where the hot vapor stream is contacted with a cooled reflux stream or pumparound stream. Before the liquid and hot vapor reach thermal equilibrium, some of the metal surfaces are cooled enough to condense water.The location of this corrosion is usually one or two trays in the tower below the naphtha reflux or the top pumparound. • Under Deposit Corrosion — This corrosion is associated with the deposition of ammonium, amine, or acidic iron chloride above the dew point of water.The area of attack often is in hot, dry exchangers or in the tower at temperatures from 20°F/11°C to 50°F/28°C above the estimated water dew point. In addition, any isolated area with limited access to vapor or liquid flow might have this type of corrosion. • Initial Condensation Corrosion — With this corrosion, a high level of corrosion is associated with the first condensation of water as the overhead vapor is normally condensed in a crude overhead system.The area of attack is usually in the top exchanger(s) of a crude unit overhead.

56

HYDROCHLORIC ACID (LOW pH — STRONG ACID) CORROSION Figure V-1 shows an example of hydrochloric acid and hydrogen sulfide corrosion.This sample of a carbon steel overhead line came from a horizontal section of the overhead system.The top half of the line had severe metal loss in the form of pitting and generalized corrosion.The bottom half of the line showed no corrosion.

Figure V-1: Hydrochloric Acid and Hydrogen Sulfide Corrosion

Closer inspection of the pits shows smooth depressions with gently sloping sidewalls and smooth bottoms (Figure V-2).

Figure V-2: Closer Inspection of the Pits

57

Possible causes of corrosion in these caps are hydrogen sulfide and ammonium chloride deposits (Figure V-3 and Figure V-4).

Figure V-3: 13 Cr Stainless Steel (UNS410) Bubble Cap Corrosion

Figure V-4: Close-up of Pitting Corrosion

58

Figure V-5 shows an example of ammonium chloride underdeposit corrosion.

Figure V-5: Ammonium Chloride Corrosion on Carbon Steel Heat Exchanger Tubes

The pits with narrow openings and deep cavenous voids are typical of hydrochloric acid pitting (Figure V-6).

Figure V-6: Pits with Narrow Openings and Deep Cavernous Voids

59

NAPHTHENIC ACID CORROSION Figure V-7 and Figure V-8 show SS410 (13Cr) tray cap parts that are severely deteriorated by ammonium chloride, and the hydrogen sulfide that formed the metal sulfides found on the surface after the alloy had been degraded.

Figure V-7: Ammonium Chloride Corrosion on SS410 Trays

Figure V-8: Ammonium Chloride Corrosion on SS410 Trays

60

304L Stainless steel tray-stressed corrosion cracking failure. Corrosion was pitting attack attributed to hydrogen sulfides or chlorides.

Figure V-9: 304L Stainless Steel Tray

Figure V-10: Stress Corrosion Cracking Failure

61

Figure V-11 and V-12 is an example of underdeposit corrosion by chlorides, sulfides, and ammonia.

Figure V-11: Carbon Steel Trays — HCl Corrosion

Figure V-12: Underdeposit Corrosion — Note the thinning cleaned surface

62

Figure V-13 shows a carbon steel upper tray with an orange peel corrosion pattern. Deposits indicate corrosion mechanisms could have been sulfidation, oxidation, or ammonium chloride.

Figure V-13: Sulfidation/Oxidation or Ammonium Chloride

63

Figure V-14 shows an example of a carbon steel heat exchanger tube in pre-flash overhead with external underdeposit corrosion.

Figure V-14. Underdeposit Corrosion

In Figure V-15, pitting and gouging are most likely from acid penetration of fouled surface. Note the internal surface showed no corrosion.

Figure V-15: Pitting and Gouging - External Surface

64

VI. NAPHTHENIC ACID CORROSION CHARACTERISTICS Form of Corrosion Can be pitting, gouging, or grooving Corrosion Products Iron naphthenates, possible iron sulfide Corrosion Reactions Fe + 2R=COOH
Fe (OOC-R)2 + H2 FeS + 2R-COOH
Fe (OOC-R)2 + H2S FeO + 2R-COOH
Fe (OOC-R)2 + H2O (not as common) Fe (OOC-R)2 + H2S
FeS + RCOOH (regenerated naphthenic acid) Unique Forms of Corrosion Usually found in areas of velocity or turbulence, and at vapor and liquid interface Treatment Improve metallurgy Add high temperature stable chemical corrosion inhibitors specifically designed for protection against naphthenic acid Examples of Corrosion Locations Crude units Vacuum units Lube oil extraction units

65

NAPHTHENIC ACID CORROSION Figure VI-1 shows an example of Naphthenic Acid attack on a 14Cr Tray.

Figure VI-1: Naphthenic Acid attack on 14Cr Tray

Upgrade in metallurgy or corrosion inhibitor required to resist naphthenic acid corrosion (Figure VI-2).

Figure VI-2: Typical Pitting From Naphthenic Acid Corrosion on Heater Plugs

66

VII. HIGH pH CORROSION CHARACTERISTICS Primarily caused by H2S Can be accelerated by cyanides in some systems Form of Corrosion Pitting, blistering, and cracking Corrosion Products FeS Corrosion Reactions (High pH) 2H2S + 2NH3 
2NH4+ + 2HS- (bisulfide ion) 2NH4+ + 2HS- + Fe 
FeS + H2 + (NH4)2S Fe + 2HS-
FeS + S-2 + 2H0 2H0
H2 FeS + 6CN-
Fe(CN)-46 + S-2 (if cyanide present) Unique Forms of Corrosion Sulfide Stress Cracking (SSC) Sulfide Stress Corrosion Cracking (SSCC) Hydrogen Induced Cracking (HIC) Stress-Oriented Hydrogen-Induced Cracking (SOHIC) Treatment Dilute with water wash Apply filming amines/corrosion inhibitors (filming amines — a specific type of the more generically used corrosion inhibitors) Use polysulfide to remove cyanides when present Apply metal passivation chemistry (amine units) Examples of Corrosion Locations FCC and FCC gas plants Cokers Sour water strippers Hydrocrackers Hydro-desulfurization (HDS)

67

Figure VII-1 and Figure VII-2 show an example of hydrogen sulfide attack on 90Cu-10Ni tubes leading to failure.

Figure VII-1: De-alloying Tubes — H2S Corrosion 90-10 Cu:Ni Heat Exchanger Tubes

Figure VII-2: Close-up of De-alloying Tubes — H2S Corrosion 90-10 Cu:Ni Heat Exchanger Tubes

68

VIII. MISCELLANEOUS EXAMPLES OF CORROSION Erosion—Corrosion Overhead line tee section Overhead exchanger tubes Amine unit

Storage Tank Corrosion Crude Preheat Corrosion Product Pipeline Corrosion Aluminum Brass Stress Cracking UAN Corrosion Fuel Storage Tank Corrosion

69

MISCELLANEOUS EXAMPLES OF CORROSION Figure VIII-1 shows an example of a carbon steel tee showing signs of erosion corrosion.The horseshoe-shaped pits indicate the direction of flow.

Figure VIII-1: Erosion Corrosion — Carbon Steel Tee

Figure VIII-2 shows a close-up of the pits after cleaning. Corrosion by-products in the pits suggest the corrosive agents were ammonium chloride and hydrogen sulfide.

Figure VIII-2: Close-up Pits after Cleaning

70

In this example of erosion-corrosion, fluids are channeled through the perforations in the diffuser plate, causing localized impingement sites. FeS film is continually eroded off, replaced, and eroded off again (Figure VIII-3).

Figure VIII-3: Erosion-Corrosion — Carbon Steel Overhead Exchanger Tubes

71

EROSION — CORROSION IN A DEA STRIPPER Metal loss was confined to a DEA impingement zone. Corrosion was attributed to carbon dioxide in the DEA (Figure VIII-4).

Figure VIII-4: Erosion-Corrosion — DEA Stripper Tower Wall

Figure VIII-5 shows the large areas that were eroded, leaving a smooth surface.

Figure VIII-5: Large Areas Eroded

72

CRUDE STORAGE TANK — CORROSION SRB's and iron oxidizing bacteria attack carbon steel storage tank bottom floor (Figure VIII-6).

Figure VIII-6: Crude Storage Tank Failure Most Likely from Wet Oxygenated Soils

73

CRUDE PREHEAT CORROSION Figure VIII-7 shows an example of a carbon steel exchanger tube from a crude pre-heat train (tubeside: crude oil, shell side: hot naphthas). • Pits were round with smooth side walls and etched bottoms. • Corrosion by-products were iron oxides, iron sulfides, and iron carbonates, suggesting 02, H2S, and CO2 as the corrosive species. • Internal deposits contained calcium, magnesium, chlorine, sodium, and silicon — probably from water in the crude.

Figure VIII-7: Oxygen, H2S, and CO2 Corrosion

74

PRODUCT PIPELINE CORROSION Underdeposit etching was attributed to sulfur containing species and oxygenated fuel in transported fuels (Figure VIII-8 and Figure VIII-9).

Figure VIII-8: Carbon Steel Product Pipeline Corrosion

Figure VIII-9: Close-up of Carbon Steel Product Pipeline Corrosion

75

ALUMINUM BRASS STRESS CRACKING Aluminum brass crude overhead heat exchanger tubes show signs of stress cracking (Figure VIII-10 and Figure VIII-11). There was no evidence of corrosion affecting the tubes.Therefore this was probably due to: • Tubes overstressed during installation. • Thermal stressing during hydrotest and operations. • Operational stresses, such as water, hammering, and vibration.

Figure VIII-10: Aluminum Brass Stress Cracking

Figure VIII-11: Aluminum Brass Stress Cracking

76

UAN CORROSION Figure VIII-12 and Figure VIII-13 shows an example of UAN corrosion. Note how the corrosion is more severe at the tank edges and along the plate boundaries.

Figure VIII-12: UAN Corrosion — Carbon Steel Storage Tank

Figure VIII-13: Close-up View of the Corrosion at a Plate Boundary

77

Figure VIII-14 shows an example of corrosion on a carbon steel storage tank bottom from concentrated UAN

Figure VIII-14: UAN Corrosion on Carbon Steel Storage Tanks

78

FUEL STORAGE TANK CORROSION Figure VIII-15 and Figure VIII-16 shows an example of corrosion by pitting at the water layer in the bottom of diesel steel storage tanks. Possible mechanisms for pitting include microorganism influenced corrosion and/or galvanic corrosion due to sulfide inclusions in the steel.

Figure VIII-15: Fuel Storage Tank Corrosion Corrosion fatigue is a combination of corrosion and cyclic stresses (Figure I-52).

Figure VIII-16: Fuel Storage Tank Corrosion

79

Figure VIII-17 is a photomicrograph at 100X showing the cavernous pits, carbides, and manganese sulfide stringers.

Figure VIII-17: Fuel Storage Tank Corrosion

Figure VIII-18 shows another section of the bottom of the fuel tank where corrosion had not occurred. Note the relative cleanliness as compared to figure VIII-17.

Figure VIII-18: Fuel Storage Tank Corrosion

80

IX.WATERSIDE FAILURES IN THE REFINING INDUSTRY Steam Generating Systems Steam generating systems include conventional fired boilers and boilers that generate steam from hot process streams such as transfer line exchangers. Several different failure mechanisms may occur on boiler tubing used in such systems. Case histories describing a few of the possible mechanisms are presented below. More detailed descriptions, and additional boiler system failure mechanisms, are presented in The Nalco Guide to Boiler Failure Analysis. Case History #1 — Failure of a Steam Generator Tube Due to Caustic Corrosion System — Waste heat steam generator (transfer line exchanger) Specimen — Steam generator tube Orientation — Vertical Material — Carbon steel Time in service — 4 years System age — 20 years Environment Internal — Quenching gases Temperature — 1,000°-1,500°F (538° - 816°C) Pressure — 75 psi External — Water and steam mixture Temperature — 650°F (343°C) Pressure — 1,500 psi Chemical treatment — Phosphate/caustic Deep, localized wastage occurred at the ends of several tubes at the bottom of a transfer line exchanger, causing two failures (Figure IX-1).Thick layers of deposits accumulated on the bottom tubesheet of this exchanger (Figures IX-2 and Figure IX-3).These deposits insulated the tubes and tubesheet, and prevented the rinsing away of caustic and other aggressive alkaline salts following steam generation.The alkaline salts concentrated to levels as high as the percent ranges to remove protective iron oxide layers and to attack the steel.

Figure IX-1: Localized Metal Loss at End of Tube that Contacted Bottom Tubesheet

81

Figure IX-2: Transfer Line Exchanger

Figure IX-3: Tubesheet where Corrosion Occurred

82

Reactions 4NaOH + Fe3O4
2NaFeO2 + Na2FeO2 + 2H2O Fe + 2NaOH
Na2FeO2 + H2 Alkaline deposits, containing sodium, covered the surface in the corroded area (Figure IX-4). Metallographic analysis revealed the presence of magnetite needles and metallic copper on the corroded surface.

Figure IX-4: Corrosion Products in Wasted Area Containing Magnetite Needles and Metallic Copper Particles Corrosion may also occur on tubes along the top tubesheet in vertical steam generators due to the formation and entrapment of a stable steam blanket.Water boils away to leave behind dissolved solids and a concentrated solution of caustic, alkaline, or even acidic salts. Corrosion Control Measures Clean and maintain clean tubesheet surfaces. Alter the design and operation of the exchanger to prevent the formation and entrapment of a steam blanket. Since no corrosion occurred away from the tube ends, boiler water chemistry and treatment was proper and did not contribute to the corrosion.

83

Case History #2 — Idle Time Oxygen Corrosion in a Power Boiler System — Power generating boiler Specimen — Riser tube Orientation — Vertical Material — Carbon steel Time in service — 30 years System age — 30 years Environment Internal — Boiler water Pressure — 850 psi Chemical treatment — Polymer, organic oxygen scavenger External — Gas turbine exhaust with supplementary refinery or natural gas as needed Temperature — 2,500°F (1371°C) maximum furnace temperature Chemical treatment — None A tube section was removed from the generating bank for surveillance.The internal surface of the section was generally covered with thin, uniform layers of black iron oxides and deposits (Figure IX-5).The uniform layers were interrupted in scattered sites by mounds or tubercles of black iron oxide (Figure IX-6).The presence of iron oxide tubercles indicated that oxygen corrosion occurred at the tubercle sites, resulting from exposure to waters containing excessive concentrations of dissolved oxygen. Removing the layers and tubercles revealed a mostly smooth, unattacked internal surface that was interrupted by moderately deep depressions where the tubercles were present (Figure IX-7).

Figure IX-5: Bottom-half Section Showing Internal Surface Covered with Uniform Layer of Black Iron Oxide Interrupted by Scattered Iron Oxide Tubercles

84

Figure IX-6: Iron Oxide Tubercle with No Flow Orientation

Figure IX-7: Depressions Formed at Tubercle Sites Revealed by Cleaning

85

The presence of black iron oxide in most areas indicated that oxygen concentration was low during boiler operation. In watertube boilers containing steam drums, it is extremely rare for significant oxygen corrosion to occur within the generating bank during boiler operation. Oxygen introduced to the drum in the feedwater tends to flash off with the steam in the drum. In addition, the scattered nature and small population of the tubercles indicated that oxygen corrosion did not occur during boiler operation.This evidence, and the lack of flow orientation of the tubercles, indicated that oxygen corrosion occurred during idle periods when water was present inside the tubes and was oxygenated by in-leaked air. Orange, brown or red are the colors of hematite, an iron oxide with a higher oxidation state than the black magnetite that normally forms on surfaces when boiler water contains the proper low dissolved oxygen concentration. Hematite formed during idle periods may be subsequently overlaid with or even reduced to a black magnetite during subsequent boiler operation. In some cases, idle time oxygen corrosion may be identified by the pattern and nature of iron oxide formation. Figure IX-8 shows the internal surface of a horizontally-oriented economizer tube that experienced shallow metal loss within narrow strips along the bottom. During idle periods the boiler was drained and a shallow pool of water remained along the tube bottom. Air leaked into the boiler and oxygenated the residual water to cause oxygen corrosion that produced orange-brown iron oxide corrosion products.

Figure IX-8: Localized Strip of Orange-brown Iron Oxide along bottom of Horizontal Tube Reaction 2Fe + H2O + O2 ♦ Fe2O3 (hematite) + 2H

86

Corrosion Control Idle time oxygen corrosion may be controlled through use of proper lay-up practice.This may include inert gas (nitrogen) blanketing and the proper use of oxygen scavenging chemicals. Such chemicals may be fed to higher levels than used during boiler operation, and may require monitoring, replenishment and circulation. Oxygen corrosion during boiler operation requires proper feedwater deaeration and oxygen scavenging practice. Case History #3 — Corrosion Fatigue Cracking of a Steam Generator Tube System — Unspecified boiler Specimen — Waterwall tube on the side wall, 10 tubes from the front burner Orientation — Vertical Material — Carbon steel Time in service — 12 years System age — 12 years Environment Internal — Water and steam mixture Temperature — 500°F (260°C) Pressure — 600 psi Chemical treatment — Polymer/sulfite External — Unspecified fuel Temperature —1,300°- 1,475°F (704° - 802°C) Multiple leaks occurred in tubes located near the received section of tubing. Regions containing families of moderately deep fissures were present on the hot side internal surface beneath thick deposit layers (Figures IX-9 and Figure IX-10). The fissures propagated transgranularly through the microstructure and were filled with dense oxide containing a centerline crack (Figure IX-11). These features indicated the fissures were caused by corrosion fatigue cracking.

Figure IX-9: Heavy Deposit Layers Covering Internal Surface of Boiler Tube

87

Figure IX-10: Numerous Transversely Oriented Fissures on Internal Surface

Figure IX-11: Fissure Caused by Corrosion Fatigue Cracking Filled with Dense Oxide Containing Centerline Crack

88

Corrosion fatigue cracking of boiler tubing results from the application of cyclic tensile stresses; the mechanism can be described as follows. Initially, the surfaces of the metal are covered with a thin layer of dense oxide (Figure IX-12a). Sufficient tensile stresses will result in fracture of the protective iron oxide layers covering the surfaces (Figure IX-12b). Where the oxide layer fractures, bare metal is exposed which immediately oxidizes (Figure IX-12c). Under subsequent stress cycles, oxide fracture and reformation occurs repeatedly at the same location to allow the cracks to grow (Figure IX-12d). No specific corrodents are required for corrosion fatigue to occur, as the protective iron oxide layer will form by reaction with water and/or steam. Although corrosion fatigue may be accelerated by oxygen corrosion, dissolved oxygen in the boiler water is not required for cracking to occur.

Figure IX-12: Illustration of the Mechanism of Corrosion Fatigue Cracking (a) Normally Protective Oxide Layer on the Metal Surface (b) Application of a Tensile Stress Cracks the Oxide Layer (c) Exposed Metal at the Tip of the Crack Oxidizes (d) Repeated, Cycling Tensile Stresses Cause the Fissure to Grow

Cyclic tensile stresses can result from a number of sources that include varying bending or tensile loads, vibrations, or temperature or pressure changes.The transverse orientation of the fissures indicated that cyclic bending stresses were responsible for cracking, as the fissures propagate in a direction perpendicular to the principle stresses.The fissures were only present on the internal surface, indicating that localized temperature changes related to the formation and collapse of a semi-stable steam blanket were likely responsible for cracking. Changing metal temperatures caused contraction and expansion of the tube metal, generating cyclic stresses. A small amount of sodium was identified in the internal surface deposits, indicating that evaporation to dryness occurred.The fissures were localized to an area slightly displaced from the hot side tube crown, likely the side facing toward the burner where heat input would be greatest.

89

Cyclic operation of boilers, including frequent start-ups and shut-downs or peaking service, can also result in corrosion fatigue cracking. Cycling changes in internal pressure of the system would tend to produce corrosion fatigue fissures having a longitudinal orientation if the resulting hoop stresses are sufficient to cause cracking. Control Measures The morphology of the fissures produced by corrosion fatigue cracking depends on numerous factors. Blunt, wedgeshaped cracks are generally associated with cyclic stresses of low frequency and/or stress level. Conversely, tight, narrow cracks are typically produced by cyclic stresses of high frequency and/or stress levels. A large number of fissures may also suggest high-level stresses. Crack arrest points, or nodes, present along the fissures indicate different periods of crack growth interrupted by corrosive attack. Stresses may concentrate at locations of physical constraint and surface discontinuities, such as depressions and notches. The fissures were localized to shallow depressions caused by old oxygen corrosion that occurred during boiler idle time (Figure IX-13).These shallow depressions acted as stress concentrators, promoting nucleation sites for fissures.

Figure IX-13: Numerous Fissures Initiating from Shallow Depressions Caused by Old Idle Time Oxygen Corrosion Preventing the steam blanketing condition responsible for cracking would eliminate corrosion fatigue cracking.This may be accomplished by chemical cleaning to remove the insulating internal surface deposits and ensuring adequate coolant flow for the heat input. In general, eliminating or reducing corrosion fatigue cracking is accomplished by identifying and decreasing the cyclic stresses responsible for cracking.

90

Case History #4 — Failure of a Boiler Deaerator Due to Stress Corrosion Cracking System — Deaerator Specimen — Dome section from degasifier Orientation — Horizontal Material — 304L stainless steel Time in service — 4 years System age — 4 years Environment Internal — Wet steam, possible splashing of returned condensate Temperature — 250°F (121°C) Pressure — 14 psi Chemical treatment — None to the deaerator and degasifier section External — Ambient air, surface covered with insulation Numerous fissures and cracks formed in the dome of the degasifier section of a deaerator, requiring replacement of the entire dome (Figure IX-14).The system was shut down for four weeks before failure, during which time the deaerator was kept hot with steam feed. No failures occurred when the unit was restarted. However, after four days of operation several cracks penetrated through the dome.The internal surface was also covered with substantial amounts of iron oxide corrosion products that formed in place.These corrosion products overlaid shallow depressions (Figure IX-15).

Figure IX-14: Cracks on Internal Surface Covered with Iron Oxide Corrosion Products

91

Figure IX-15: Iron Oxide Corrosion Products Removed to Reveal Depressions Metallographic cross-sections revealed that the cracks were transgranular, highly branched and penetrated from the internal surface outward (Figure IX-16). This revealed that the fissures and cracks formed due to stress corrosion cracking.

Figure IX-16: Metallographic Cross-section Revealing Branched, Transgranular Cracks Penetrating Outward from Internal Surface

92

Stress corrosion cracking results from exposure of a particular alloy to specific corrodents. Stresses that are required to cause cracking may be applied, or may be residual stresses within the metal that are supplied by original metal forming operations and welding. The possible corrodents that may have caused cracking in a deaerator include chloride, and caustic or other alkaline salts. Neither of these agents should have been present in the degasifier section. However, contamination of the steam, condensate or feedwater may have occurred, possibly on an intermittent basis. Corrodents may have concentrated to corrosive levels due to evaporation. No residue of either chloride or alkaline salts was found on the internal surface. This is not unusual, since such agents are very water soluble and would wash away after exposure when exposed to normal, noncontaminated water. The most probable corrodent responsible for stress corrosion cracking was determined to be chloride for several reasons. 1) The threshold concentration of chloride necessary to cause cracking may be quite low at the normal operating temperatures of the deaerator. Cracking may occur at chloride concentration as low as 20 ppm at 200°F (93°C). 2) Caustic stress corrosion cracking rarely occurs at metal temperatures below 300°F (149°C) and generally requires concentrations of caustic in the percent levels. It is improbable that either condition was met in this system. 3) The presence of depressions and oxide corrosion products on the internal surface would have formed due to corrosion by chloride. Caustic would not tend to cause such corrosion. Control of Stress Corrosion Cracking Since chloride stress corrosion cracking may occur at even low concentrations at normal operating temperatures, it is recommended that deaerators and water-wetted components throughout boilers should not be constructed of austenitic stainless steels. Aside from this recommendation, chloride stress corrosion cracking may be prevented only by ensuring that no chloride gets into the deaerator, which will be very difficult and expensive to accomplish. Case History #5 — Overheating in a Power Boiler System — Power generating boiler Specimen — Rear wall tube Orientation — Vertical Material — Carbon steel Time in service — 26 years System age — 26 years Environment Internal — Boiler water Pressure — 1,200 psi Chemical treatment — Equilibrium phosphate, organic oxygen scavenger External — Natural gas fired Temperature — 2,200°F (1204°C) maximum furnace temperature Chemical Treatment — None

93

Two bulges formed within a localized area on the hot side of a boiler wall tube (Figure IX-17).The bulges occurred in a tube that was located near a burner. Internal deposition was light in all areas except at the bulges (Figure IX-18).

Figure IX-17: Localized Bulges on Hot Side of Wall Tube

Figure IX-18: Localized Deposition at Bulge Site

94

Metallographic analysis indicated that the original pearlite microstructure of the tube was maintained in all locations except the bulges (Figure IX-19). At the bulges, the microstructure changed to scattered spheroidal carbides in a ferrite matrix (Figure IX-20). Creep fissures also formed on the bulge external surfaces (Figure IX-21). Moderately thick layers of thermally-formed iron oxide built up on both internal and external surfaces at the bulges.

Figure IX-19: Microstructure of Tube in As-manufactured Condition in All Locations Except at Bulges Consisting of Colonies of Pearlite in Ferrite Matrix

Figure IX-20: Microstructure at Bulges Consisting of Spheroidal Carbides in Ferrite Matrix

95

Figure IX-21: Intergranular Creep Fissures on External Surface at Bulge All of the microstructural features at the bulges indicated that the tube wall at the bulge sites experienced mild, longterm overheating. Metal temperatures ranged from about 850° to 1,100°F (454° to 593°C) for a period ranging from several weeks to several months. Metals have decreased strength at high temperatures (Figure IX-22). Consequently, the tube wall bulged under stresses supplied by internal pressure.The localized nature of the bulging and internal deposition indicated that overheating was caused by excessive fireside heat input.

Figure IX-22: Metal Strength versus Temperature (Source: The Nalco Guide to Boiler Failure Analysis)

96

Control of Overheating The most practical method of controlling overheating in this case would be through proper insulation of the external surfaces of the tubes in the location where the bulges formed.This may be provided by installing refractory or repairing damaged refractory at that location. If this cannot be accomplished, then excessive heat input should be prevented by decreasing firing rates and preventing localized channeling of hot gases to the zone of the bulged tube. Overheating of boiler tubes may be caused by several different factors, including not only excessive heat input, but also internal fluid flow rate and deposition, possibly in combination.Therefore, the cause of any bulging or rupturing failure of a boiler tube should be diagnosed with a thorough metallographic examination. Case History #6 — Failure of a CO Boiler Economizer Tube Due to Fireside Corrosion System — Carbon monoxide boiler Specimen — Economizer tube Orientation — Horizontal Material — Corten steel (carbon steel containing about 0.2% copper) Time in service — 11 years System age — 30 years Environment Internal — Boiler feedwater Temperature — 275°F (135°C) inlet, 390°— 415°F (199° — 213°C) outlet Pressure — 950 psi Chemical Treatment — Chelant, polymer, sulfite External — Carbon monoxide from fluid catalytic cracker, refinery gas, steam from soot-blowing Temperature — 450°F (232°C) Pressure — 5 — 8 psi Chemical treatment — None Numerous deep depressions formed on the external surfaces of several different economizer tubes (Figure IX-23).Thick layers and mounds of light-colored deposits covered the depressions and surfaces in the surrounding areas. Deep depressions were found beneath adherent deposits (Figure IX-24).These deposits tested acidic in a distilled water solution. Chemical analysis of the deposits revealed the presence of high concentrations of sulfate and volatile sulfur compounds.The presence of these deposits indicates that the fireside corrosion may have been caused in part by the condensation of sulfuric acid on the surfaces.When high sulfur content fuels are combusted, the sulfur is oxidized to sulfur trioxide, which combines with moisture in the furnace environment to form sulfuric acid.

97

Figure IX-23: External Surface Depressions and Acidic, Light-colored Deposits

Figure IX-24: Deposits Removed to Reveal Depressions

Reactions S + O2
SO2 SO2 + O
SO3, SO2 + 1/2 O2 + catalyst
SO3 H2O + SO3
H2SO4 H2SO4 + Fe
FeSO4 + H2

98

The dewpoint of sulfuric acid varies as a function of moisture and sulfur trioxide concentration (Figure IX-25 and Figure IX-26). Significant amounts of moisture were supplied to the furnace environment by nearby soot-blowers that emitted steam to remove fireside deposits from boiler tubes.Typically, the operating temperature of the economizer was well above the sulfuric acid dewpoint at any water and sulfur trioxide concentration.Therefore, corrosion during boiler operation was improbable. However, acidic deposits remaining on surfaces during idle periods became hydrated by condensed water to form acids that caused corrosion.

Figure IX-25: Sulfuric Acid Dewpoint as a Function of Fuel Sulfur Content (Source: The Nalco Guide to Boiler Failure Analysis)

Figure IX-26: Sulfuric Acid Dewpoint as a Function of Water Vapor Concentration in Furnace Environment (Source: The Nalco Guide to Boiler Failure Analysis) Corrosion Control Measures If the fuel cannot be changed to one with a lower sulfur concentration, a few corrosion control measures may be effective.

99

Fireside acid corrosion during boiler operation may be controlled, if practical, by: 1) 2) 3) 4)

Maintaining metal temperatures above the sulfuric acid dewpoint. Reducing moisture content (lower moisture content fuels and minimization of soot-blowing). Altering firing practice, by combusting fuel with lower excess air. Feeding fuel treatment additives.

Corrosion by acidic fireside deposits during idle periods may be controlled by: 1) Removing corrosive acidic fireside deposits with high pressure water sprays followed by rinsing with a neutralizing lime wash spray. 2) Maintaining a dry fireside environment with the aid of dessicants. Corten steel, normally used to control atmospheric corrosion, was chosen to help control idle time corrosion. However, this alloy was inadequately resistant to the aggressive, acidic fireside deposits.

100

Cooling Water Systems Cooling water systems include overhead condensers, surface condensers and a wide variety of process fluid heat exchangers. Several different corrosion and other failure mechanisms may occur on the watersides of heat exchanger tubing used in such systems. Some of these mechanisms are highlighted in the following case histories. More detailed descriptions, and additional cooling water system failure mechanisms are presented in The Nalco Guide to Cooling Water Systems Failure Analysis. Case History #7 — Failure of a Heat Exchanger Tube Due to Stress Corrosion Cracking System — Depentanizer Specimen — Heat exchanger tube (externally finned) Orientation — Horizontal Material — Admiralty brass Time in service — 10 years System age — 40 years Environment Internal — Cooling water Temperature - 85°-110°F (29° - 43°C) Pressure — 45 —70 psi Chemical treatment — Ortho/organic/polyphosphate, tolyltriazole, polyacrylate dispersant External — Crude unit light naphtha (C2 — C4) Temperature - 110°-120°F (43° — 49°C) Pressure — 45 psi Approximately 200 heat exchanger tubes within the unit experienced sudden cracking failures (Figure IX-27). Metallographic analysis revealed the presence of deep, highly branched, intergranular cracks that penetrated from the internal, waterside surface (Figure IX-28).This mode of crack propagation is characteristic of stress corrosion cracking. Stress corrosion cracking of metals results from exposure to specific corrodents for the particular alloy while being subjected to applied and/or residual stresses. Residual stresses are supplied by original tube forming operations.

Figure IX-27: Wall of Admirality Brass Heat Exchanger Tube Penetrated by a Crack

101

Figure IX-28: Branched, Intergranular Crack and Fissures Penetrating from Internal Surface The specific corrodents that most commonly cause stress corrosion cracking of admiralty brass are ammonia and hydrogen sulfide.These corrodents were considered as possible causes for the cracking. However, analysis of deposit and corrosion product layers that covered the internal surface revealed the presence of significant amounts of mercury compounds. Mercury causes rapid intergranular cracking of copper alloys in a manner similar to stress corrosion cracking, a mechanism referred to as liquid metal embrittlement.The rapidity of the failures and the intergranular propagation of the cracks more significantly implicate mercury as the corrodent.The presence of mercury is highly unusual in cooling water and could only have been introduced due to contamination. One possible source of mercury is from a pressure-measuring device. Control Measures Corrodents that may cause stress corrosion cracking of admiralty brass at low concentrations, such as mercury, should be eliminated from the environment. Other corrodents, such as ammonia and hydrogen sulfide, should be kept to low concentrations to avoid cracking. Reduction of stresses may control cracking to a less significant degree.

102

Case History #8 — Stress Corrosion Cracking of a Brass Heat Exchanger Tube System — Propane condenser Specimen — Exchanger tube from the first pass of the condenser Orientation — Horizontal Material — Inhibited admiralty brass Time in service — Unknown System age — Unknown Environment Internal — Cooling water Temperature — 65°- 80°F (18° - 27°C) Pressure — 35 psi Chemical Treatment — Phosphonate/tolytriazole/chlorine, pH 8.0 External — Propane vapor Temperature — 150°F (66°C) Pressure — 135 psi Numerous leaks occurred in the system over a short period of time.The received section contained a thick wall fracture on one end (Figure IX-29). Multiple crack initiation sites were present along the circumference. Numerous tight, transversely oriented cracks and fissures were present on the internal surface near the fractured end (Figure IX-30). Stratified deposit layers covered a smooth surface contour that showed no significant evidence of corrosion on the internal surface.

Figure IX-29: Brittle Fracture Edge Containing Multiple Crack Initiation Sites

103

Figure IX-30: Numerous Transversely Oriented Fissures near Fracture Edge

Metallographic cross sections revealed many tight, highly branched fissures indicative of Stress Corrosion Cracking (SCC) (Figure IX-31).The cracks propagated transgranularly into the tube microstructure (Figure IX-32). SCC of brass may be transgranular or intergranular. SCC of a susceptible alloy requires the simultaneous presence of sufficient tensile stress and a specific corrodent.The stress and corrodent interact synergistically to produce cracks on the surfaces exposed to the corrodent and propagate in response to the stress state.The transverse orientation of the cracks and multiple initiation sites along the circumference indicated that the tensile stresses responsible for cracking were uniaxial (longitudinally oriented).The tensile stresses required for cracking can be applied or residual. Residual stresses can be developed from fabrication and installation of the tubes.

104

Figure IX-31: Tight Branched Fissure Caused by Stress Corrosion Cracking

Figure IX-32: Transgranular Propagation Path of a Fissure

105

Stress corrosion cracking in copper alloys occurs most frequently in environments of aqueous solutions or moist atmospheres that contain ammonia or amines. Sulfur dioxide, hydrogen sulfide, nitrates and nitrites have also been shown to produce stress corrosion cracking in copper-base alloys.The fissures on the section initiated on the internal surface. It was reported that the cooling water was contaminated with amines for a period of time just prior to leaks developing in this condenser. Control Measures Stress corrosion cracking can be prevented by removing one of the requirements of cracking, (i.e., tensile stress, the corrodent or the susceptible alloy). In general, the propensity and severity of the cracking increases as stress levels rise. For many alloys, there is a threshold stress below which stress corrosion cracking will not occur (Figure IX-33). Threshold stress depends on many factors, including temperature, alloy composition and concentration of the corrodent. Reduction of the applied and/or residual stresses can reduce or eliminate cracking. Preventing contamination into the system of corrodents responsible for stress corrosion cracking will eliminate cracking. In addition, changing to a less susceptible material may also prevent cracking. Substitution of 70-30 cupronickel in place of admiralty or other copper-base alloys has been suggested as a preventative measure for stress corrosion cracking in refinery service.

Figure IX-33: Increase in Time to Failure with Decreasing Stress and Threshold Stress Below Which Cracking Does Not Occur

106

Case History #9 — Failure of a Heat Exchanger Tube due to Dealloying and Corrosion Fatigue Cracking System — Surface condenser Specimen — Condenser tube Orientation — Horizontal Material — Admiralty brass Time in service — 6 years System age — 20 years Environment Internal — Cooling water (high chloride concentration, > 400 ppm) Temperature — 100°- 118°F (38°- 48°C) Pressure — 5 kgf Chemical treatment — Organic phosphate, dispersant, bleach External — Steam condensate from a boiler Temperature — 130°- 140°F (55°- 60°C) Pressure — Slight vacuum A crack-like failure occurred in a surface condenser tube (Figure IX-34). Scattered depressions formed on the internal surface (Figure IX-35). Metallographic analysis revealed that the surfaces of the depressions were covered with layers of metallic copper, indicating that dezincification, a form of dealloying, occurred (Figure IX-36). Dezincification occurred beneath occlusive deposit layers due to the concentration of chloride from the water solution by an ion concentration cell mechanism.The corrosion caused zinc to be preferentially removed from the copper-zinc matrix of the alloy.This left behind a weak, porous mass of pure copper on the corroded surface.

Figure IX-34: Cracking Failure

107

Figure IX-35: Internal Surface Depressions

Figure IX-36: Metallic Copper Covering Surface, Indicating Dezincification Failure occurred primarily due to corrosion fatigue cracking.The crack initiated at the base of one of these depressions, which served as a point of stress concentration (Figure IX-37).This cracking resulted from the application of cyclic tensile stresses.These stresses were supplied by cyclic bending, expansion and contraction, and vibration in service.

108

Figure IX-37: Corrosion Fatigue Fissures Penetrating from Base of Depression

Corrosion Control Dezincification may be controlled by maintaining clean metal surfaces, and by avoiding low or stagnant flow. Exposure to acidic or alkaline waters, or high chloride waters, promotes dezincification.Therefore, the use of an alternate, more resistant alloy such as cupro-nickel may be considered. Corrosion fatigue cracking may be controlled most effectively by minimizing the magnitude and frequency of the applied stresses. Case History #10 — Failure of a Heat Exchanger Tube due to Microbiologically Influenced Corrosion System — Lean amine cooler Specimen — Heat exchanger tube Orientation — Horizontal Material — Carbon steel Time in service — 2 1/2 years System age — 7 years Environment Internal — Lean amine Temperature — 140°F (60°C) Pressure — 2,800 kPa External — Cooling water Temperature — 112°F (45°C) Pressure — 650 kPa Chemical treatment — Zinc phosphate, polymer, nonoxidizing biocide

109

Clusters of mutually intersecting, deep depressions formed in scattered spots on the external surfaces of several tubes in the exchanger (Figure IX-38).The depressions were hemispherical in shape. In some locations, the depressions mutually intersected (Figure IX-39).The depressions were lined with iron sulfide corrosion products (Figure IX-40). It was reported that a gooey slime layer covered the external surface when the bundle was removed. Bacterial counts on tube surfaces were 5 million CFU/ml total, with 8,000 CFU/ml of sulfate reducing bacteria.

Figure IX-38: Scattered, Deep External Surface Depressions

Figure IX-39: Clusters of Mutually-intersecting, Rounded Depressions

110

Figure IX-40: Metallographic Cross-section of Depression Covered with Corrosion Products Containing Sulfide

In any failure analysis investigation, all four factors listed below are required to positively confirm a case of microbiologically influenced corrosion. In this case, all four factors were identified to confirm that corrosion was caused by sulfate reducing bacteria. 1) 2) 3) 4)

The presence of microorganisms — Confirmed by high bacteria counts. Unique corrosion morphology — Hemispherical depressions, some in clusters. Specific corrosion products and deposits — Sulfide layers lining the depressions. Compatible environmental conditions — Sulfate reducing bacteria are anaerobic.

Sulfate reducing bacteria reduce sulfate ions to sulfide in their biological process. One of the more popular proposed mechanisms to explain how these bacteria promote corrosion is through removal of hydrogen, which depolarizes the cathode in the corrosion cell. Sulfate reducing bacteria are anaerobic, so they live and grow beneath occlusive layers of deposits and slime. Slime also promotes the growth of anaerobic bacteria by depleting oxygen. Corrosion Control Microbiologically influenced corrosion may be controlled through use of adequate concentrations of proper biocides that are fed in a proper manner. Low (< 3 ft/s) or stagnant (<1 ft/s) flow should be minimized or prevented. Clean metal surfaces should also be maintained. In this case history, flow rates were low, since cooling water flowed across the shell sides of the tubes. In addition, slug feed of biocide was inadequate for the system.

111

Case History # 11 — Failure of a Heat Exchanger Tube due to Concentration Cell Corrosion System — Splitter bottoms cooler Specimen — Heat exchanger tube Orientation — Horizontal Material — Carbon steel Time in service — 2 years System age — 25 years Environment Internal — Cooling water Temperature — 75°F (24°C) in, 95°F (35°C) out Chemical treatment — Organic phosphate, bleach External — Heavy lean oil Temperature — 317°F (158°C) in, 90°F (32°C) out Pressure — 12 psi The internal surface of a steel heat exchanger tube section was covered with thin layers of light-colored deposits that consisted primarily of calcium phosphate (Figure IX-41).These deposits overlaid layers and tubercles of brown iron oxide. Some of the tubercles were hollow and lined with material that was acidic and contained significant amounts of chloride and sulfur compounds.The hollow tubercles were broken away to reveal rough-surfaced, undercut depressions that were bare, or lined with very thin deposit and corrosion product layers.These surfaces could be cleaned to bare metal quickly with a concentrated hydrochloric acid solution (Figure IX-42). Metallographic cross-sections also revealed undercut depressions that were capped with hollow tubercles (Figure IX-43).

Figure IX-41: Iron Oxide Layers and Tubercles Overlaid with Water-formed Deposit Layers

112

Figure IX-42: Deposit and Corrosion Product Layers Removed to Reveal Rough-surfaced, Undercut Depressions

Figure IX-43: Metallographic Cross-section of Rough-surfaced, Undercut Depressions beneath Hollow Iron Oxide Tubercle

113

Initial corrosion was caused by the formation of differential oxygen concentration cells beneath occlusive deposit layers. A diagram of a differential oxygen concentration cell between contacting metal surfaces is shown in Figure IX-44.The water beneath an occlusive deposit or corrosion product layer or within a tight crevice between contacting metal surfaces becomes depleted in oxygen concentration. Consequently, the cathodic reaction in the corrosion cell could only occur where oxygen concentrations were higher, such as at breaks in the deposit layer, and on the outer surfaces of iron oxide layers and tubercles.The anodic reaction in the corrosion cell occurs within the oxygen depleted zone to produce metal dissolution. Metal ion leaks out of the oxygen depleted zone into the zone where oxygen concentration is high to produce iron oxide corrosion products. By this process, iron oxide tubercles form.

Figure IX-44: (Source: The Nalco Guide to Cooling Water Systems Failure Analysis)

114

Reactions Oxygen concentration cell Anodic reaction Fe
Fe2+ + 2eCathodic reaction

O2

Corrosion products formed

+ 2H2O + 4e-
4OH2 Fe2+ + 4e- + O2 + 2H2O
2Fe(OH)2

Ion concentration cells also may form beneath occlusive deposit and corrosion product layers, tubercles and between crevices (Figure IX-45).The positively charged metal ions within the occluded zones attract aggressive anions such as chloride, sulfate, and sulfide. Aggressive anion concentration may become several times that in the bulk water to produce locally acidic and highly corrosive environments.The acidic environments dissolve protective corrosion product layers and then the metal itself to produce rough-surfaced, undercut pits and depressions. Acid corrosion significantly accelerates metal loss.The reactions and corrosion products that may occur within an iron oxide tubercle are shown in Figure IX-46.

Figure IX-45: (Source: The Nalco Guide to Cooling Water Systems Failure Analysis)

115

Figure IX-46: Tubercle (Source: The Nalco Guide to Cooling Water Systems Failure Analysis)

Corrosion Control Measures Concentration cell corrosion may be controlled by maintaining clean metal surfaces, both mechanically and through proper water treatment and corrosion inhibition. Corrosion may also be controlled by minimizing or eliminating low or stagnant flow, and by decreasing concentration of aggressive anions. If these measures cannot be taken, and/or if corrosion cannot be controlled, then the use of alternate materials of coatings may be considered.

116

ABOUT THE AUTHORS William (Bill) Tillis Consultant William (Bill) Tillis has 46 years experience in the field of corrosion. Presently retired from Exxon, he is consulting with Nalco on oilfield and refinery failure analysis. He has worked for Curtis-Wright, Pennwalt, FMC Corp., General Electric, NL Industries, NL Treating Chemicals, and Exxon Chemicals. Bill has worked the past 23 years in oilfield corrosion inhibitor testing and failure analysis. Bill received a BS degree in chemistry from Long Island University, and a MS in chemistry from Drexel University. He is a NACE certified corrosion specialist and an active member of both NACE and the American Society of Metals (ASM). Bill holds several U.S. patents and has authored numerous SPE and NACE papers.

Bill Tillis

117

James (Jim) Dillon Research Scientist Jim has been a part of the Nalco metallurgy group for 20 years, serving as senior research metallurgist, staff scientist and his current position. He has conducted more than 3,000 failure analysis investigations on components from a variety of industrial systems, both in the laboratory and during on-site inspections. Jim has been involved in the consultation and education of customers as well as sales, marketing and research personnel concerning the identification and mitigation of corrosion and other failure mechanisms. Jim received a BS degree in metallurgical engineering from the Illinois Institute of Technology. His professional associations include the American Society for Metals (ASM), National Association of Corrosion Engineers (NACE), Association of Iron and Steel Engineers (AISE), and American Iron and Steel Institute (AISI). Jim has authored numerous articles for NACE.

Paul Desch Senior Research Metallurgist Paul has been with Nalco for 6 years serving as a senior metallurgist and senior research metallurgist. Paul has extensive experience in boiler systems and cooling water systems, especially regarding material selection and failure modes associated with different materials. He currently serves as the lead contact for the company's compatibility testing program, which encompasses many aspects of product compatibility with metals, nonmetals and coatings. Paul received a BS degree in 1988 and a PhD in 1996 from the Illinois Institute of Technology in Chicago. His professional associations include the American Society for Metals (ASM), National Association of Corrosion Engineers (NACE), and the Society of Plastics Engineers (SPE).

Paul Desch (Left) and Jim Dillon (Right)

118

Energy Services Division Worldwide Headquarters P.O. Box 87, Sugar Land,Texas 77487-0087 U.S.A. +1 281 263 7000 SUBSIDIARIES AND AFFILIATES IN PRINCIPAL LOCATIONS AROUND THE WORLD www.nalco.com ©2004

Nalco Company All Rights Reserved Printed in USA 11/04 NE-66 DS/OFC